Merge, Split, Extract or Delete: Choosing the Right PDF Page Operation
Four tools that look interchangeable and are not. What each one actually does to the file, which document structures survive, and the order to run them in so you do not lose work.
Merge, Split, Extract Pages and Delete Pages are the four operations almost every PDF task decomposes into. They overlap enough that people pick whichever one they used last, and then spend ten minutes undoing the result.
The differences are small in the interface and significant in the output. Here is what each actually does.
The one thing they have in common
All four work the same way underneath: a new, empty PDF is created, and the pages you want are copied into it. Nothing is edited in place. Your source file on disk is never modified — every operation writes a new document.
This has two consequences worth internalising. First, mistakes are cheap: the original is always still there. Second, page content survives perfectly — text stays vector text, images keep their original encoding, embedded fonts come along — but document-level structures do not. Bookmarks and outlines, form fields, and some annotation types live at the document level, not on individual pages, so they are left behind when pages are copied into a new container.
Merge: many documents, one file
Merge takes several PDFs and produces one, in the order you set by dragging the queue. Use it for assembling a submission pack, combining scanned batches, or putting a cover letter in front of a report.
Two practical notes. The merged file is roughly the sum of the inputs — merging is not a compression step, and nothing is re-encoded. And encrypted PDFs are rejected outright rather than skipped silently, so a password-protected input has to be unlocked in the application that produced it first.
Split: one document, many files
Split is the inverse. You give it ranges — "1-5, 6-10, 11" — and each range becomes its own output file, delivered individually or as a ZIP.
Reach for it when the outputs have separate lives: one chapter per file, one invoice per file, one client's section per file. If everything you are keeping is going to the same person anyway, Extract is the better fit.
Extract: keep these pages
Extract Pages gathers your selection into a single new document. "Send me clauses 4 to 9" is an Extract job; "give me each chapter separately" is a Split job. That is the whole distinction.
One detail catches people out: printed page numbers are drawn onto the pages themselves, so extracting pages 40 to 45 produces a document whose first page still reads "40". Run Add Page Numbers afterwards if you need a fresh sequence.
Delete: remove these pages
Delete Pages produces the same class of result as Extract, but you describe the pages to discard instead of the pages to keep. On a 200-page scan with six blank separator sheets, that is the difference between listing six numbers and listing a hundred and ninety-four.
Because the output is assembled page by page, deleted pages are genuinely absent from the new file — they are not hidden in a layer that a determined reader could recover. That is not the same as redaction, though: blacking out a paragraph on a page you keep requires a dedicated redaction tool, because text under a drawn rectangle is still text.
Order matters
When you are chaining operations, sequence saves work:
- Remove pages before merging. Cleaning each source first is easier than hunting through a combined 300-page document.
- Number pages last. Any page operation after numbering leaves a broken sequence printed on the pages.
- Watermark last as well. Same reason — and it means you only stamp the pages that survived.
- Compress last of all. Compression renders pages to images, so anything you do afterwards is working on pictures rather than text.
Quick reference
- Several files in, one file out → Merge
- One file in, several files out → Split
- One file in, one smaller file out, easier to say what stays → Extract
- One file in, one smaller file out, easier to say what goes → Delete
All four run entirely inside your browser — which matters more than usual here, because the documents people restructure are typically contracts, court bundles, medical records and tender submissions. The file never leaves the tab.
Related reading
Continue reading
Your PDF Is Too Big to Email. Here Is What Actually Works
Attachment limits are smaller than people assume, and most "compress PDF" advice quietly destroys your text layer. A walkthrough of what makes a PDF large, which fix suits which document, and what each one costs you.
What Your Photos Say About You: EXIF, GPS and When to Strip It
Every photo carries a hidden record: where it was taken, when, on what device, sometimes down to the serial number. Here is what is actually in there, who can read it, and the simplest ways to remove it.